Biases in probability judgments: hot hand versus gambler's fallacy

Abstract

This study focuses on how people's causal beliefs about processes that lead to a sequence of events influence their probability judgments on the next outcome. These beliefs affect people's tendencies towards either of the two well-known prediction biases: the hot hand and the gambler's fallacy. In two experiments, we examined the effect of processes that include human performance to people's predictions, probability judgments and randomness judgments. It was observed that when the process that generates a sequence of events includes human performance, the degree of control that the agent, whose performance generates the sequence, has over the outcomes significantly affects subject's randomness judgments about that process, but not their predictions on the next outcome of that sequence of events. When the sequence is generated by a random mechanism, subjects preferred gambler's fallacy for their prediction strategies. However, this effect can be overridden by the effect of the structure of the sequence (alternating vs. streaky) given to the subjects, which also influenced subject's predictions and probability judgments. Suggestions for further research are discussed.